[PATCH v2 01/18] arm64: topology: Use __free(put_cpufreq_policy) for policy reference
Zihuan Zhang
zhangzihuan at kylinos.cn
Thu Aug 28 12:32:46 AEST 2025
在 2025/8/27 17:17, Sudeep Holla 写道:
> On Wed, Aug 27, 2025 at 09:30:13AM +0100, Ben Horgan wrote:
>> Hi Zihuan,
>>
>> On 8/27/25 03:31, Zihuan Zhang wrote:
>>> Replace the manual cpufreq_cpu_put() with __free(put_cpufreq_policy)
>>> annotation for policy references. This reduces the risk of reference
>>> counting mistakes and aligns the code with the latest kernel style.
>>>
>>> No functional change intended.
>>>
>>> Signed-off-by: Zihuan Zhang <zhangzihuan at kylinos.cn>
>>> ---
>>> arch/arm64/kernel/topology.c | 9 +++------
>>> 1 file changed, 3 insertions(+), 6 deletions(-)
>>>
>>> diff --git a/arch/arm64/kernel/topology.c b/arch/arm64/kernel/topology.c
>>> index 5d07ee85bdae..e3cb6d54f35b 100644
>>> --- a/arch/arm64/kernel/topology.c
>>> +++ b/arch/arm64/kernel/topology.c
>>> @@ -307,17 +307,16 @@ int arch_freq_get_on_cpu(int cpu)
>>> */
>>> if (!housekeeping_cpu(cpu, HK_TYPE_TICK) ||
>>> time_is_before_jiffies(last_update + msecs_to_jiffies(AMU_SAMPLE_EXP_MS))) {
>>> - struct cpufreq_policy *policy = cpufreq_cpu_get(cpu);
>>> + struct cpufreq_policy *policy __free(put_cpufreq_policy);
>> Based on the guidance, in include/linux/cleanup.h, I would expect the
>> assignment to be done on this line.
>>
>> "...the recommendation is to always define and assign variables in one
>> * statement and not group variable definitions at the top of the
>> * function when __free() is used."
>>
> Agreed. I did something similar recently and there was a code path where
> variable wasn't initialised and ended up with freeing unassigned pointer.
> So it is more than just a recommendation sometimes.
>
Thanks a lot for your suggestions.
We are also considering introducing a WITH_CPUFREQ_POLICY wrapper to
encapsulate the cpufreq_cpu_get/put usage, so that the release order
won’t be accidentally changed.
Link:
https://lore.kernel.org/all/874d821e-8ea3-40ac-921b-c19bb380a456@kylinos.cn/
Do you have any suggestions or preferences on this direction?
More information about the Linuxppc-dev
mailing list